MR Volumes Research Articles

Differentiation between intra-axial metastatic tumor progression and radiation injury following fractionated radiation therapy or stereotactic radiosurgery using MR spectroscopy, perfusion MR imaging or volume progression modeling

Objective To determine the accuracy of magnetic resonance spectroscopy (MRS), perfusion MR imaging (MRP), or volume modeling in distinguishing tumor progression from radiation injury following radiotherapy for brain metastasis. Methods Twenty-six patients with 33 intra-axial metastatic lesions who underwent MRS ( n=41) with or without MRP ( n=32) after cranial irradiation were retrospectively studied. The final diagnosis was based on histopathology ( n=4) or magnetic resonance imaging (MRI) follow-up with clinical correlation ( n=29). Cho/Cr (choline/creatinine), Cho/NAA (choline/ N-acetylaspartate), Cho/nCho (choline/contralateral normal brain choline) ratios were retrospectively calculated for the multi-voxel MRS. Relative cerebral blood volume (rCBV), relative peak height (rPH) and percentage of signal-intensity recovery (PSR) were also retrospectively derived for the MRPs. Tumor volumes were determined using manual segmentation method and analyzed using different volume progression modeling. Different ratios or models were tested and plotted on the receiver operating characteristic curve (ROC), with their performances quantified as area under the ROC curve (AUC). MRI follow-up time was calculated from the date of initial radiotherapy until the last MRI or the last MRI before surgical diagnosis. Results Median MRI follow-up was 16 months (range: 2–33). Thirty percent of lesions ( n=10) were determined to be radiation injury; 70% ( n=23) were determined to be tumor progression. For the MRS, Cho/nCho had the best performance (AUC of 0.612), and Cho/nCho >1.2 had 33% sensitivity and 100% specificity in predicting tumor progression. For the MRP, rCBV had the best performance (AUC of 0.802), and rCBV >2 had 56% sensitivity and 100% specificity. The best volume model was percent increase (AUC of 0.891); 65% tumor volume increase had 100% sensitivity and 80% specificity. Conclusion Cho/nCho of MRS, rCBV of MRP, and percent increase of MRI volume modeling provide the best discrimination of intra-axial metastatic tumor progression from radiation injury for their respective modalities. Cho/nCho and rCBV appear to have high specificities but low sensitivities. In contrast, percent volume increase of 65% can be a highly sensitive and moderately specific predictor for tumor progression after radiotherapy. Future incorporation of 65% volume increase as a pretest selection criterion may compensate for the low sensitivities of MRS and MRP.

A brain MRI study of chronic fatigue syndrome: evidence of brainstem dysfunction and altered homeostasis

To explore brain involvement in chronic fatigue syndrome (CFS), the statistical parametric mapping of brain MR images has been extended to voxel-based regressions against clinical scores. Using SPM5 we performed voxel-based morphometry (VBM) and analysed T1- and T2-weighted spin-echo MR signal levels in 25 CFS subjects and 25 normal controls (NC). Clinical scores included CFS fatigue duration, a score based on the 10 most common CFS symptoms, the Bell score, the hospital anxiety and depression scale (HADS) anxiety and depression, and hemodynamic parameters from 24-h blood pressure monitoring. We also performed group × hemodynamic score interaction regressions to detect locations where MR regressions were opposite for CFS and NC, thereby indicating abnormality in the CFS group. In the midbrain, white matter volume was observed to decrease with increasing fatigue duration. For T1-weighted MR and white matter volume, group × hemodynamic score interactions were detected in the brainstem [strongest in midbrain grey matter (GM)], deep prefrontal white matter (WM), the caudal basal pons and hypothalamus. A strong correlation in CFS between brainstem GM volume and pulse pressure suggested impaired cerebrovascular autoregulation. It can be argued that at least some of these changes could arise from astrocyte dysfunction. These results are consistent with an insult to the midbrain at fatigue onset that affects multiple feedback control loops to suppress cerebral motor and cognitive activity and disrupt local CNS homeostasis, including resetting of some elements of the autonomic nervous system (ANS). © 2011 The Authors. NMR in Biomedicine published by John Wiley & Sons, Ltd.

Accuracy of image registration between MRI and light microscopy in the ex vivo brain

A multistep procedure was developed to register magnetic resonance imaging (MRI) and histological data from the same sample in the light microscopy image space, with the ultimate goal of allowing quantitative comparisons of the two datasets. The fixed brain of an owl monkey was used to develop and test the procedure. In addition to the MRI and histological data, photographic images of the brain tissue block acquired during sectioning were assembled into a blockface volume to provide an intermediate step for the overall registration process. The MR volume was first registered to the blockface volume using a combination of linear and nonlinear registration, and two dimensional (2D) blockface sections were registered to corresponding myelin-stained sections using a combination of linear and nonlinear registration. Before this 2D registration, two major types of tissue distortions were corrected: tissue tearing and independent movement of different parts of the brain, both introduced during histological processing of the sections. The correction procedure utilized a 2D method to close tissue tears and a multiple iterative closest point (ICP) algorithm to reposition separate pieces of tissue in the image. The accuracy of the overall MR to micrograph registration procedure was assessed by measuring the distance between registered landmarks chosen in the MR image space and the corresponding landmarks chosen in the micrograph space. The average error distance of the MR data registered to micrograph data was 0.324±0.277 mm, only 8% larger than the width of the MRI voxel (0.3 mm).

Employing Symmetry Features for Automatic Misalignment Correction in Neuroimages

A novel method to automatically compute the symmetry plane and to correct the 3D orientation of neuro-images is presented. In acquisition of neuroimaging scans, the lack of perfect alignment of a patient's head makes it challenging to evaluate brain images. By deploying a shape-based criterion, the symmetry plane is defined as a plane that best matches external surface points on one side of the head, with their counterparts on the other side. In our method, the head volume is represented as a re-parameterized surface point cloud, where each location is parameterized by its elevation (latitude), azimuth (longitude), and radius. The search for the best matching surfaces is implemented in a multi-resolution paradigm, and the computation time is significantly decreased. The algorithm was quantitatively evaluated using in both simulated data and in real T1, T2, Flair magnetic resonance patient images. This algorithm is found to be fast (<10s per MR volume), robust and accurate (<.6 degree of Mean Angular Error), invariant to the acquisition noise, slice thickness, bias field, and pathological asymmetries.

Robust automatic rigid registration of MRI and X‐ray using external fiducial markers for XFM‐guided interventional procedures

In X-ray fused with MRI, previously gathered roadmap MRI volume images are overlaid on live X-ray fluoroscopy images to help guide the clinician during an interventional procedure. The incorporation of MRI data allows for the visualization of soft tissue that is poorly visualized under X-ray. The widespread clinical use of this technique will require fully automating as many components as possible. While previous use of this method has required time-consuming manual intervention to register the two modalities, in this article, the authors present a fully automatic rigid-body registration method. External fiducial markers that are visible under these two complimentary imaging modalities were used to register the X-ray images with the roadmap MR images. The method has three components: (a) The identification of the 3D locations of the markers from a full 3D MR volume, (b) the identification of the 3D locations of the markers from a small number of 2D X-ray fluoroscopy images, and (c) finding the rigid-body transformation that registers the two point sets in the two modalities. For part (a), the localization of the markers from MR data, the MR volume image was thresholded, connected voxels were segmented and labeled, and the centroids of the connected components were computed. For part (b), the X-ray projection images, produced by an image intensifier, were first corrected for distortions. Binary mask images of the markers were created from the distortion-corrected X-ray projection images by applying edge detection, pattern recognition, and image morphological operations. The markers were localized in the X-ray frame using an iterative backprojection-based method which segments voxels in the volume of interest, discards false positives based on the previously computed edge-detected projections, and calculates the locations of the true markers as the centroids of the clusters of voxels that remain. For part (c), a variant of the iterative closest point method was used to find correspondences between and register the two sets of points computed from MR and X-ray data. This knowledge of the correspondence between the two point sets was used to refine, first, the X-ray marker localization and then the total rigid-body registration between modalities. The rigid-body registration was used to overlay the roadmap MR image onto the X-ray fluoroscopy projections. In 35 separate experiments, the markers were correctly registered to each other in 100% of the cases. When half the number of X-ray projections was used (10 X-ray projections instead of 20), the markers were correctly registered in all 35 experiments. The method was also successful in all 35 experiments when the number of markers was (retrospectively) halved (from 16 to 8). The target registration error was computed in a phantom experiment to be less than 2.4 mm. In two in vivo experiments, targets (interventional devices with pointlike metallic structures) inside the heart were successfully registered between the two modalities. The method presented can be used to automatically register a roadmap MR image to X-ray fluoroscopy using fiducial markers and as few as ten X-ray projections.

Prognostic significance of parameters derived from co-registered18F-fluorodeoxyglucose PET and contrast-enhanced MRI in patients with high-grade glioma

The aim of this study was to determine the prognostic significance of the volume and intensity of abnormal (18)F-fluorodeoxyglucose positron emission tomography (FDG-PET) accumulation within areas of contrast enhancement on post-therapeutic volumetric MRI. A total of 10 patients with Grade III or IV glioma were treated with resection followed by intracavitary radiation therapy with (131)I-labelled antitenascin monoclonal antibody. Patients underwent serial FDG-PET and 1.5 T MR imaging. For each patient, MR and FDG-PET image volumes at each time point were aligned using a rigid-body normalised mutual information algorithm. Contrast-enhancing regions of interest (ROIs) were defined using a semi-automated k-means clustering technique. Activity within the ROI on the co-registered PET scan was calculated as a ratio (mean activity ratio; MAR) to activity in contralateral normal-appearing white matter (NAWM). The PET lesion was defined as the portion of the ROI associated with activity greater than two standard deviations above the mean in NAWM. Survival was assessed using the logrank test. Larger contrast-enhancing ROIs were strongly associated with an increased MAR (r = 0.51; p<0.002). Enhancing lesions with an MAR >1.2 were associated with decreased survival (p<0.016). In nine patients who died, the MAR on PET correlated inversely with survival duration (r = -0.43; p<0.01), whereas PET lesion volume did not. Following intracavitary radiation therapy, the development of contrast-enhancing lesions that are associated with high mean FDG-PET accumulation suggests poor prognosis.

SU‐GG‐T‐177: Intensity‐Modulated Radiotherapy Planning Characteristics of Celiac Deep‐Seated Tumor

Objective: To examine the influence of energy and number of beams for intensity‐modulated radiation therapy on celiac deep‐seated tumor. Methods and Materials: 26 patients with cancer of pancreas were studied. Their IMRT plans were created for 6MV and 15MV photons using 5, 7, 9 and 11 coplanar non‐opposed fields. Plans were normalized to cover no less than 95% of the target volume at the 100% prescribed dose and were analyzed using two factor ANOVA with repetition and broken line graph for (a) dose distribution on PTV, (b) conformity index at prescribed dose for PTV, (c) dose on the organs at risk(OAR), (d) dose distribution on the region around the target which were divided into “near region”(NR: 1‐cm thick shell surrounding target), “surface region”(SR: 1cm thick shell under the surface) and “middle region”(MR: the region except NR and SR). Result: Dose distribution on PTV and OAR were the same for all the eight different plans; CI for all the other six plans had no significant deviations except the two 5‐beams plans which had less CI; Plans with the same number of beams but lower beam energy would enlarge the MR and SR volume significantly; Plans with 7 or 9 beams showed the better characteristics with the same energy. Conclusion: With IMRT for the celiac deep‐seated tumor, plans with 6MV and 15MV photons had no significant deviations on target and OAR, although the lower energy would enlarge the dose volume on MR and SR. The clinical significance of this increased dose, in terms of complications, remains to be detected. Considering the disadvantages of using the higher energy including higher secondary radiation dose to the patient and staff, increased non‐equilibrium effects and larger penumbra, the plan with 6MV photon and 7 beams should be considered first.

Abstract LB-382: MR microscopy detects architectural changes in the pancreas of mutant Kras mouse models of pancreatic cancer

Abstract Pancreatic cancer (PC) is a devastating disease due to the lack of early detection resulting in late diagnosis and poor prognosis. Thus, there is a need for a technology that can visualize morphological and cellular changes in the pancreas when the cancer cells are at a pre-invasive stage. Nearly 90% of human PC have a mutation in Kras and several mouse models that express mutant Kras develop precancerous lesions resembling those in humans. Using MR microimaging, we examined pancreata from EL-Kras (ELK) and Pdx1-Cre/LSL-Kras (PCK) mice which target mutant Kras to the pancreas via unique approaches. EL targeting employs human mutant Kras where nearly all acinar cells express this oncogene, while Pdx1 targeting uses a mutation in an endogenous mouse Kras allele. The source (human vs. mouse), targeting (EL vs. Pdx1) and delivery (artificial vs. endogenous) of mutant Kras contribute to altered phenotypes including differences in normal appearing parenchyma, frequency of acinar-ductal metaplasia, precancerous lesion formation, and progression to advanced disease. We examined pancreas from both mutant Kras mouse models using MR microimaging and compared their MR signature to that of control mice. Fixed pancreata were imaged on a 14.1T microimager. T2-weighted, high spatial resolution images were acquired with fat suppression using a fast spin-echo 3D imaging protocol and isotropic pixel size of ∼70μm. MR data was reconstructed to provide volume images of the pancreatic components and their spatial relationship to each other. T2-weighted MR microimages had high spatial resolution and excellent endogenous contrast that enabled the visualization of acinar cells, islets and stroma in the normal pancreas. T2-weighted images of ELK pancreas showed poor contrast and delineation between acinar cells, islets and stroma indicating changes in their T2-relaxation times possibly as a result of changes at the cellular level. Contrast between pancreatic components in the PCK pancreas was similar to control. Volume images of the pancreas from both mutant Kras mice showed unique differences from control in stromal architecture. While the normal mouse pancreas had a dense, concentrated network of stroma, the ELK mouse pancreas showed only a sparse network with reduced signal intensity, and the PCK mouse pancreas had a more extensive network of stroma. These differences in stromal architecture of mutant Kras pancreas detected by MR microscopy most likely represent transgene-induced alterations in the composition of extracellular matrix (ECM) and might have implications for the development of pancreatic pre-cancerous lesions and cancer. Ours is the first report of MR microimaging of ECM in the intact mouse pancreas of relevant mouse models of PC. Deciphering the differences in ECM among these two models is the focus of ongoing work. Our results suggest that MR volume imaging at high spatial resolution may have potential in detecting morphological and cellular changes associated with pre-invasive and invasive stages of PC. (Supported by Jay Wertheimer Pancreatic Cancer Research Fund and NIH S10 RR13880) Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 101st Annual Meeting of the American Association for Cancer Research; 2010 Apr 17-21; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2010;70(8 Suppl):Abstract nr LB-382.

Modeling brain tissue volumes over the lifespan: quantitative analysis of postmortem weights and in vivo MR images

Normative measurements of brain gray matter and white matter tissue volumes across the lifespan have not yet been established. The purpose of this article was to use mathematical modeling and analytical functions to demonstrate the growth trajectory of gray matter and white matter from age 0 to age 90. For each gender, brain weight functions were generated by utilizing existing autopsy data from 4400 subjects. Brain gray matter, white matter and lateral ventricular volumes were measured from 39 MR volumes of normal individuals. These were converted to weight by multiplying the tissue volumes by the specific gravity of that tissue. White matter volumes were described by a saturating exponential function, and the gray matter volume function was calculated by subtracting the white matter weight function from the brain weight function. For each gender, equations were generated for white matter and gray matter volumes as a function of age over the lifespan.

MEG Detects Alpha-Power Modulations in Pulvinar

Event Abstract Back to Event MEG Detects Alpha-Power Modulations in Pulvinar Yohan Attal1*, Jérôme Yelnik2, Eric Bardinet1, Marie Chupin1 and Sylvain Baillet3 1 Massachusetts General Hospital, United States 2 Hopital Pitie-Salpetriere, INSERM U679, France 3 Medical College of Wisconsin, United States Objectives: Neuroimaging evidence regarding the functional organization of subcortical nuclei such as pulvinar is rather scarce compared to cortical structures [1]. Further, MEG is often believed to be insensitive to deeper brain structures. We have recently developed a global-brain MEG model that encompasses structural and electrophysiological properties of cortical and subcortical structures for source analysis [2]. Here we use this model to map the brain structures involved in the modulation of brain activity in the alpha range in 6 subjects. Methods and Results: Individual T1-weighted MRI axial scans (1x1x1.5mm voxel size) from 6 healthy subjects were processed to extract cortical and subcortical tessellations. Neocortical and hippocampus surface tessellations were obtained using the automatic pipeline available in the BrainVISA free software solutions [3]. Central grey nuclei were segmented from the individual MR volumes using a deformable atlas approach [4]. Elementary current dipoles (ECD) were placed at the surface of the cortex and hippocampus and along regular volumic grids within subcortical nuclei. The MEG session consisted of a block-designed “eyes open - eyes closed” paradigm (CTF Omega whole-head 151-channel MEG system). Each block lasted for 30 s. Data were band-pass filtered within the alpha band that was determined on an individual basis after inspection of the power spectra of each subject. Distributed source imaging was performed using BrainStorm. Significant ( , non-parametric permutation tests [5]) modulations of MEG source amplitudes were detected by contrasting the eyes open vs. eyes closed conditions. These amplitude modulations were detected in the dorsal and lateral parietal cortex and in the pulvinar thalamic regions. These areas are densely interconnected and generally assumed to be involved in the generation and modulation of alpha brain waves. Conclusions: The global source model has been evaluated on experimental MEG data on a classical eyes-open/eyes-closed paradigm. Modulations of activities were detected in parieto-occipital cortical regions and in the pulvinar part of the thalamus. These results indicate that complete, realistic modelling of MEG source distribution can help detect activation in deeper brain structures that were thought to be invisible to MEG.

Feasibility of Functional Cardiac MR Imaging in Mice Using A Clinical 3 Tesla Whole Body Scanner

To test the feasibility of cardiac MR imaging in mice using a clinical 3 Tesla whole body MR system for structural and functional analysis. Standard protocols for bright blood cine imaging were adapted for murine dimensions. To validate measurements of functional parameters the MR data were compared with high-resolution echocardiographic measurements. Cardiac imaging was carried out in CD 1 wild-type mice (n = 8). MR imaging studies were performed using a clinical 3 Tesla MR system (Achieva, Philips). All mice received 2 MR scans and 1 echocardiographic evaluation. For optimal MR signal detection a dedicated solenoid receive-only coil was used. Electrocardiogram signal was recorded using a dedicated small animal electrocardiogram monitoring unit. For imaging we used a retrospectively triggered TFE sequence with a repetition time of 12 ms and an echo time of 4 ms. A dedicated software patch allowed for triggering of cardiac frequency of up to 600 BPM. Doppler-echocardiography was performed using a VisualSonics Vevo 770 high-resolution imaging system with a 30 MHz scanhead. Axial/lateral resolution was 40 of 100 microm and temporal resolution was 150 to 300 frames/s (B-mode) and 1000 frames/s (M-mode) depending on the setting. MR imaging was successfully carried out in all mice with a sufficient temporal resolution and good signal-to-noise ratio and contrast-to-noise ratio levels allowing for identification of all relevant structures. Accordingly, there was a good scan-rescan reproducibility of MR measurements: Interassay coefficients of variance ranged from 4% for ejection fraction to 12% for endsystolic volume (ESV). Magnetic resonance imaging and echocardiography gave comparable results when using the same geometric model (Teichholz method): EDV: 60.2 +/- 6.1 microL/59.1 +/- 12.3 microL, ESV: 20.0 +/- 2.6 microL/20.7 +/- 7.7 microL, EF: 66.7% +/- 4.0%/65.2% +/- 9.9%, CO 19.5 +/- 3.6 mL/17.9 +/- 2.9 mL. Bland-Altman analysis gave acceptable limits of agreement between both methods: EDV (+28.2/-26.1), ESV (+16.3/-17.7), EF (+19.0/-16.1), CO (10.7/-7.5). When applying the Simpson's method MR volume estimates were significantly higher compared with echocardiography resulting in a lower estimate for the ejection fraction (60% +/- 3.9% vs. 66.7% +/- 4.0%). Cardiac MR imaging of mice using a clinical 3 Tesla MR system for functional analysis is feasible with sufficient spatial and temporal resolution, good repeatability and reliable results when compared with high-resolution echocardiography.

Early metabolic responses in temozolomide treated low-grade glioma patients

Amino acid transport and protein synthesis are important steps of tumor growth. We investigated the time course of tumor metabolism in low-grade gliomas (LGG) during temozolomide chemotherapy, and compared metabolic responses as measured with positron emission tomography (PET) with volume responses as revealed by magnetic resonance imaging (MR). A homogeneous population of 11 patients with progressive non-enhancing LGG was prospectively studied. Imaging was done at 6-months intervals starting six months, and in a second series starting three months after treatment initiation. F-18 fluoro-ethyl-L-tyrosine (FET) uptake was quantified with PET as metabolically active tumor volume, and was compared with the tumor volume on MR. Response was defined as >or=10% reduction of the initial tumor volume. Eight patients showed metabolic responses. Already 3 months after start of chemotherapy the active FET volumes decreased in 2 patients to a mean of 44% from baseline. First MR volume responses were noted at 6 months. Responders showed a volume reduction to 31 +/- 23% (mean +/- SD) from baseline for FET, and to 73 +/- 26% for MR. The time to maximal volume reduction was 8.0 +/- 4.4 months for FET, and 15.0 +/- 3.0 months for MR. The initial metabolic response correlated with the best volume response on MR (Spearman Rank P = 0.011). Deactivation of amino acid transport represents an early indicator of chemotherapy response in LGG. Response assessment based on MR only has to be reconsidered. The time window obtained from PET may assist for individual treatment decisions in LGG patients.

Breast cancer molecular profiles and tumor response of neoadjuvant doxorubicin and paclitaxel: The I-SPY TRIAL (CALGB 150007/150012, ACRIN 6657)

LBA515 Background: I-SPY is a multi-center trial designed to identify predictive markers of pathological complete response (pCR) and survival of women with locally advanced breast cancers (3cm or greater). Women received neoadjuvant doxorubicin and cyclophosphamide then paclitaxel. Methods: 237 women enrolled, 216 completed serial imaging and core biopsies. Pre-treatment assays include: Agilent expression arrays, MIP aCGH, p53 gene chip and sequencing, IHC and reverse phase protein arrays (RPMA). Response to therapy was measured by serial MRI, pCR and residual cancer burden (RCB). Associations among molecular markers, pCR, RCB and survival were evaluated using chi-square test, Kaplan-Meier curves and log-rank test. Results: Median tumor size was 6cm, % pCR and RCB 0/1 was 27% and 36% for the entire study; % pCR rate for the 144 Agilent arrays was 25%. Distribution, rates of pCR and RCB 0/1 are shown in the Table for molecular and IHC markers. DFS and OS will be presented. Several molecular subtypes, including NKI 70 gene low, luminal A, 21 gene set low and IHC HR+, define 15–28% of patients with 3–10% pCR, yet excellent early survival. Wound healing, most discriminatory for prognosis, is not predictive of chemotherapy response. By RPMA, patients with pCR had increased phosphorylation of 4EBP1, eNOS, cAbl, STAT5, EGFR, AKT (p&lt;0.05). In ER+ patients with poor MR response, pIRS, pIGFR, p706S were activated (p&lt;0.05). RCB is a more refined way to measure pCR and was more predictive of DFS and OS (p=0.01) than pCR alone with a mean follow up of 3.9 years. MR volume is highly predictive of pCR and RCB. For specific subtypes, e.g. basal, RCB is predictive of DFS (p&lt;0.00001). Conclusions: LABC have aggressive biology. Response to therapy and outcome can be predicted by many biomarkers. The I-SPY data set provides a platform to compare, contrast and combine marker signatures to tailor therapy and demonstrates the power of the neoadjuvant setting. Support: ACRIN U01 CA079778 ; CALGB CA31964, CA33601; NCI SPORE CA58207. [Table: see text] [Table: see text]

Breast cancer molecular profiles and tumor response of neoadjuvant doxorubicin and paclitaxel: The I-SPY TRIAL (CALGB 150007/150012, ACRIN 6657)

LBA515 Background: I-SPY is a multi-center trial designed to identify predictive markers of pathological complete response (pCR) and survival of women with locally advanced breast cancers (3cm or greater). Women received neoadjuvant doxorubicin and cyclophosphamide then paclitaxel. Methods: 237 women enrolled, 216 completed serial imaging and core biopsies. Pre-treatment assays include: Agilent expression arrays, MIP aCGH, p53 gene chip and sequencing, IHC and reverse phase protein arrays (RPMA). Response to therapy was measured by serial MRI, pCR and residual cancer burden (RCB). Associations among molecular markers, pCR, RCB and survival were evaluated using chi-square test, Kaplan-Meier curves and log-rank test. Results: Median tumor size was 6cm, % pCR and RCB 0/1 was 27% and 36% for the entire study; % pCR rate for the 144 Agilent arrays was 25%. Distribution, rates of pCR and RCB 0/1 are shown in the Table for molecular and IHC markers. DFS and OS will be presented. Several molecular subtypes, including NKI 70 gene low, luminal A, 21 gene set low and IHC HR+, define 15–28% of patients with 3–10% pCR, yet excellent early survival. Wound healing, most discriminatory for prognosis, is not predictive of chemotherapy response. By RPMA, patients with pCR had increased phosphorylation of 4EBP1, eNOS, cAbl, STAT5, EGFR, AKT (p<0.05). In ER+ patients with poor MR response, pIRS, pIGFR, p706S were activated (p<0.05). RCB is a more refined way to measure pCR and was more predictive of DFS and OS (p=0.01) than pCR alone with a mean follow up of 3.9 years. MR volume is highly predictive of pCR and RCB. For specific subtypes, e.g. basal, RCB is predictive of DFS (p<0.00001). Conclusions: LABC have aggressive biology. Response to therapy and outcome can be predicted by many biomarkers. The I-SPY data set provides a platform to compare, contrast and combine marker signatures to tailor therapy and demonstrates the power of the neoadjuvant setting. Support: ACRIN U01 CA079778; CALGB CA31964, CA33601; NCI SPORE CA58207. Profile rates of pCR and RCB Distribution pCR RCB 0,1 pCR p-value RCB p-value Intrinsic Subtypes Luminal A 28% −5% −14% <0.001 <0.001 n=144 LuminalB 21% 13% −21% . . Basal 31% 34% −40% . . Her 2 15% 55% −73% . . Normal −4% 50% −67% . . NKI 70 gene Good 15% 10% −17% 0.16 0.07 n=144 Poor 85% 27% −39% . . IHC HR+ Her2− 29% 10% −21% Effect HR 0.0002 n=189 HR+ Her2+ 20% 28% −48% Effect Her2 0.009 HR− Her2− 34% 36% −41% HR− Her2− 17% 48% −79% MIP No Amplification 17q 86% 17% . n=127 Amplification 17q 14% 61% <0.0001 p53 Mutation Any Mutation 43% 38% −43% . n=181 MissenseZn-binding 39% 75% 100% 0.02 Null 27% 57% −53% 0.0003 ROR-S Low 22% −6% −21% 0.001 0.005 n=144 Medium 38% 17% −27% . . High 40% 43% −52% . . Wound Healing Quiescent 49% 19% −34% 0.12 0.72 n=144 Activated 51% 31% −38% . . 21 gene set Low 22% −3% −12% <0.001 0.002 n=144 Intermediate 10% −0% −20% . . High 68% 36% −46% . . Author Disclosure Employment or Leadership Consultant or Advisory Role Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration Agendia Agendia

Image-driven population analysis through mixture modeling.

We present iCluster, a fast and efficient algorithm that clusters a set of images while co-registering them using a parameterized, nonlinear transformation model. The output of the algorithm is a small number of template images that represent different modes in a population. This is in contrast with traditional, hypothesis-driven computational anatomy approaches that assume a single template to construct an atlas. We derive the algorithm based on a generative model of an image population as a mixture of deformable template images. We validate and explore our method in four experiments. In the first experiment, we use synthetic data to explore the behavior of the algorithm and inform a design choice on parameter settings. In the second experiment, we demonstrate the utility of having multiple atlases for the application of localizing temporal lobe brain structures in a pool of subjects that contains healthy controls and schizophrenia patients. Next, we employ iCluster to partition a data set of 415 whole brain MR volumes of subjects aged 18 through 96 years into three anatomical subgroups. Our analysis suggests that these subgroups mainly correspond to age groups. The templates reveal significant structural differences across these age groups that confirm previous findings in aging research. In the final experiment, we run iCluster on a group of 15 patients with dementia and 15 age-matched healthy controls. The algorithm produces two modes, one of which contains dementia patients only. These results suggest that the algorithm can be used to discover subpopulations that correspond to interesting structural or functional "modes."

Effects of medial temporal lobe degeneration on brain perfusion in amnestic MCI of AD type: deafferentation and functional compensation?

Cortical atrophy is correlated with the progression of neuropathological lesions within the medial temporal lobes (MTL) in Alzheimer's disease (AD). Our aim was to determine which local and remote functional changes result from MTL volume loss at the predementia stage. We studied the relationship between entorhinal and hippocampal MR volumes and whole-brain SPECT perfusion via a voxel-based correlative analysis in 19 patients with amnestic mild cognitive impairment with a memory profile suggestive of early AD. Right MTL volumes were positively correlated with remote posterior perfusion of the posterior cingulate cortex, and negatively correlated with remote anterior perfusion of the right medial and dorsolateral prefrontal cortex. There was no local correlation between volumes and perfusion within the MTL. These findings provide further insight into functional changes that result from MTL volume loss during the predementia stage of AD. The positive correlation between MTL volumes and posterior cingulate perfusion may reflect the deafferentation of a temporocingulate network due to mediotemporal degeneration. The paradoxical negative correlation between MTL volumes and prefrontal perfusion may result from recruitment of an alternative anterior temporofrontal network. It remains to be investigated how the "net sum" of this perfusion modulation affects memory and other cognitive domains through a possible compensatory perspective.

Sulcal and ventricular trajectories in stereotactic surgery

The authors analyzed deep brain stimulation electrode trajectories on MR images to identify risks of cerebrovascular complications associated with the number of electrode insertions, traversal of a sulcus, and penetration of the ventricle. Pre- and postoperative MR volumes were fused to determine the proximity of electrodes to a sulcus or ventricle and whether there were cortical, subcortical, or intraventricular complications. Complications were further classified as hemorrhagic or nonhemorrhagic and symptomatic or asymptomatic. The authors examined 258 electrode implantation for deep brain stimulation. There were 4 symptomatic events (1.6% incidence): 3 hemorrhagic and 1 nonhemorrhagic, all within the cortex. Asymptomatic events included cortical hemorrhage in 1 patient, nonhemorrhagic cortical changes in 6, pallidal hemorrhage in 1, thalamic infarction in 1, and intraventricular hemorrhage (IVH) in 5 patients. Proximity to a sulcus was a significant risk factor for hemorrhagic and nonhemorrhagic cortical complications (p = 0.001). There was a complication rate of 10.1% within the trajectories penetrating or adjacent to a sulcus, and a 0.7% rate with trajectories clearly positioned within the gyrus. Asymptomatic IVH was observed in 5% of ventricular penetrations. A history of hypertension was a risk factor for cortical hemorrhage (p = 0.019), but not for cortical ischemic/edematous events (p = 0.605). The number of electrode penetrations did not differ between patients with and without complications (p = 0.868), and the sequence of electrode insertions was not a risk factor in bilateral surgeries. Symptomatic cortical complications occur when electrodes traverse close to a sulcus. Asymptomatic IVH occurs infrequently with ventricular penetration. Despite intraoperative efforts to avoid cortical sulci, a higher than expected incidence of electrode proximity to the sulci was identified on careful postoperative trajectory analysis. This finding emphasizes the importance of assiduously planning trajectories and reviewing cases with thorough MR analysis.

Prostate delineation using CT and MRI for radiotherapy patients with bilateral hip prostheses

Background and purpose Prostate cancer patients with bilateral hip prostheses (b-HP) are rare, but pose unique challenges for high-dose radiotherapy. The purpose of this study was to determine the effect of b-HP on prostate delineation using CT and MR. Methods and materials The planning CT and MR of 7 patients with b-HP were co-registered. Four observers delineated the prostate on both data sets, blinded to the volumes delineated by their peers and those on the alternate image set. Inter-modality differences and inter-observer variability were quantified by calculating 3D volume overlaps. The effect of prostheses physical extent on these variables was evaluated. MR geometric distortion was explored using the differences in vector distances between three intra-prostatic markers (IPMs) on CT and MR. Results Prostate volumes on CT were consistently larger than MR volumes in all patients when averaged over the 4 observers (+10% to +46%, p < 0.001). The mean inter-modality volume overlap was 1.59, which varied significantly between patients (1.35 to 1.82, p = 0.04). There was a trend toward smaller inter-observer variability in the MR volumes (1.95 vs. 1.71, p = 0.08). No differences could be identified between the IPM vectors on CT and MR. Conclusions Although limited by a small sample size, this study found that MR prostate volumes were significantly smaller than those from CT for b-HP patients. The largest differences between CT and MR prostate volumes and the greatest inter-observer variability on CT tended to be seen in patients with the largest prostheses. b-HP do not appear to cause measurable MR geometric distortions in the sequences used.

Enhancement of histological volumes through averaging and their use for the analysis of magnetic resonance images

Magnetic resonance imaging (MRI) of small animals is routinely performed in research centers. But despite its many advantages, MR still suffers from limited spatial resolution which makes the interpretation and quantitative analysis of the images difficult, particularly for small structures of interest within areas of significant heterogeneity. One possibility to address this issue is to complement the MR images with histological data, which requires reconstructing 3D volumes from a series of 2D images. A number of methods have been proposed recently in the literature to address this issue, but deformation or tearing during the slicing process often produces reconstructed volumes with visible artifacts and imperfections. In this paper, we show that a possible solution to this problem is to work with several histological volumes, reconstruct each of these separately and then compute an average. The resulting histological atlas shows structures and substructures more clearly than any individual volume. We also propose an original approach to normalize intensity values across slices, a required preprocessing step when reconstructing histological volumes. We show that the histological atlas we have created can be used to localize structures and substructures, which cannot be seen easily in MR images. We also create an MR atlas that is associated with the histological atlas. We show that using the histological volumes to create the MR atlas is better than using the MR volumes only. Finally, we validate our approach quantitatively on MR image volumes by comparing volumetric measurements obtained manually and obtained automatically with our atlases.

In Vitro Assessment of a 3D Segmentation Algorithm Based on the Belief Functions Theory in Calculating Renal Volumes by MRI

Renal volumetry is an essential part of split renal function assessment in MR urography. The aim of this study was to assess the accuracy and repeatability of a 3D segmentation algorithm based on the belief functions theory for calculating renal volumes from MR images. The true volumes of 20 animal kidneys of various sizes were obtained by fluid displacement. Each kidney was examined using two different MR units. Three-dimensional proton density-weighted acquisitions with an incremental slice thickness were performed. The MR volume was then measured with a segmentation algorithm based on the belief functions theory. Two independent observers performed all segmentations twice. Accuracy, intraobserver variability, and interobserver variability were evaluated by the Bland-Altman method. The number and type of manual corrections were recorded as well as the entire processing time. The mean renal volume estimated by fluid displacement was 114 mL (range, 38-224 mL). With regard to the renal volumes obtained from assessments of adjacent axial MR images, the maximal SDs of the difference were 2.2 mL (accuracy), 0.6 mL (intraobserver variability), and 1.8 mL (interobserver variability). Segmentation of axial slices provided better accuracy and reproducibility than coronal slices. Overlapped coronal slices yielded poor results because of the partial volume effect. The mean processing time including optional manual modifications was less than 75 seconds. The belief functions theory can be considered an accurate and reproducible mathematic method to assess renal volume from MR adjacent images.